[dpdk-stable] patch 'net/mlx5: fix deadlock of link status alarm' has been queued to LTS release 16.11.5

[luca.boccassi at gmail.com]

Hi,

FYI, your patch has been queued to LTS release 16.11.5

Note it hasn't been pushed to http://dpdk.org/browse/dpdk-stable yet.
It will be pushed if I get no objections before 01/28/18. So please
shout if anyone has objections.

Thanks.

Luca Boccassi

---
>From f6f2470ea69cc55b3cd33da7e39d3d3176fb8cf0 Mon Sep 17 00:00:00 2001
From: Yongseok Koh <yskoh at mellanox.com>
Date: Wed, 10 Jan 2018 09:46:49 -0800
Subject: [PATCH] net/mlx5: fix deadlock of link status alarm

[ upstream commit 6817ad38718451849b672b0584176c14a8b55b88 ]

If mlx5_dev_link_status_handler() is executed while canceling the alarm,
deadlock can happen because rte_eal_alarm_cancel() waits for all callbackes
to finish execution and both calls are protected by priv->lock.

Fixes: 198a3c339a8f ("mlx5: handle link status interrupts")

Signed-off-by: Yongseok Koh <yskoh at mellanox.com>
Acked-by: Nelio Laranjeiro <nelio.laranjeiro at 6wind.com>
---
 drivers/net/mlx5/mlx5.h        | 16 ++++++++++++++++
 drivers/net/mlx5/mlx5_ethdev.c | 13 +++++++++----
 2 files changed, 25 insertions(+), 4 deletions(-)

diff --git a/drivers/net/mlx5/mlx5.h b/drivers/net/mlx5/mlx5.h
index 79b7a6008..0ffb94019 100644
--- a/drivers/net/mlx5/mlx5.h
+++ b/drivers/net/mlx5/mlx5.h
@@ -160,6 +160,22 @@ priv_lock(struct priv *priv)
 	rte_spinlock_lock(&priv->lock);
 }
 
+/**
+ * Try to lock private structure to protect it from concurrent access in the
+ * control path.
+ *
+ * @param priv
+ *   Pointer to private structure.
+ *
+ * @return
+ *   1 if the lock is successfully taken; 0 otherwise.
+ */
+static inline int
+priv_trylock(struct priv *priv)
+{
+	return rte_spinlock_trylock(&priv->lock);
+}
+
 /**
  * Unlock private structure.
  *
diff --git a/drivers/net/mlx5/mlx5_ethdev.c b/drivers/net/mlx5/mlx5_ethdev.c
index 2ea995f89..c7822835d 100644
--- a/drivers/net/mlx5/mlx5_ethdev.c
+++ b/drivers/net/mlx5/mlx5_ethdev.c
@@ -1239,8 +1239,12 @@ mlx5_dev_link_status_handler(void *arg)
 	struct priv *priv = dev->data->dev_private;
 	int ret;
 
-	priv_lock(priv);
-	assert(priv->pending_alarm == 1);
+	while (!priv_trylock(priv)) {
+		/* Alarm is being canceled. */
+		if (priv->pending_alarm == 0)
+			return;
+		rte_pause();
+	}
 	priv->pending_alarm = 0;
 	ret = priv_dev_link_status_handler(priv, dev);
 	priv_unlock(priv);
@@ -1287,9 +1291,10 @@ priv_dev_interrupt_handler_uninstall(struct priv *priv, struct rte_eth_dev *dev)
 	rte_intr_callback_unregister(&priv->intr_handle,
 				     mlx5_dev_interrupt_handler,
 				     dev);
-	if (priv->pending_alarm)
+	if (priv->pending_alarm) {
+		priv->pending_alarm = 0;
 		rte_eal_alarm_cancel(mlx5_dev_link_status_handler, dev);
-	priv->pending_alarm = 0;
+	}
 	priv->intr_handle.fd = 0;
 	priv->intr_handle.type = RTE_INTR_HANDLE_UNKNOWN;
 }
-- 
2.14.2